Estimating PDFs in heterogeneous radar clutter

Real-world radar resolution cells often contain several terrain types, such that the observed normalized radar cross section of a given clutter cell is an aggregate response from the underlying terrains. These heterogeneous resolution cells create an ill-posed inverse problem in estimating the scattering phenomenology from each terrain type. Assuming the terrain composition of each cell is known (e.g. through Land Use Land Cover maps) and adopting a linear mixing model, we formulate a well-posed but numerically challenging inverse problem. The Markov Chain Monte Carlo technique known as Metropolis Hasting is successfully implemented to extract the underlying Gamma clutter distribution parameters of each terrain type from the heterogeneous clutter cells.